Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A detection system configured to detect an object, the detection system comprising: a light source configured to illuminate the object; an image sensor comprising a sensor array having both infrared pixels and color pixels, the image sensor configured to output a picture; and a processor configured to generate an IR picture and a color picture according to the picture received from the image sensor, wherein the IR picture contains an IR object image and the color picture contains a color object image which corresponds to the IR object image in the IR picture, identify a skin-color object from the color object image in the color picture, and determine an object image from the IR object image in the IR picture using the identified skin-color object by comparing the identified skin-color object and the IR object image in the IR picture to remove an interference image, which does not correspond to the skin-color object of the color object image, from the IR object image.
A detection system identifies objects using a combination of infrared (IR) and color data. It includes a light source to illuminate the object, an image sensor with both IR and color pixels to capture a picture, and a processor. The processor generates an IR picture and a color picture from the sensor data. The processor identifies skin-colored areas in the color picture. It then uses this skin-color information to filter the IR picture, removing irrelevant background and interference that doesn't correspond to skin, improving object detection in the IR image.
2. The detection system of claim 1 , wherein the processor is configured to perform an AND operation between the IR picture and the color picture as comparing the identified skin-color object and the IR object image in the IR picture.
The detection system described above (light source, image sensor with IR and color pixels, processor generating IR and color pictures, skin-color identification, and IR picture filtering using skin-color data) refines the IR picture by performing an AND operation between the IR and color pictures, using the skin-color identified areas in the color picture as a mask for the IR picture to filter out interference.
3. The detection system of claim 1 , wherein the skin-color object is identified using a skin-color algorithm based on RGB color mode, HSV color mode, YUV color mode, YIQ color mode or YCbCr color mode.
In the detection system described above (light source, image sensor with IR and color pixels, processor generating IR and color pictures, skin-color identification, and IR picture filtering using skin-color data), the system identifies skin color using a skin-color detection algorithm based on standard color models such as RGB, HSV, YUV, YIQ, or YCbCr.
4. The detection system of claim 1 , wherein the processor is further configured to compare an object shape of the IR object image in the IR picture with a predetermined shape.
The detection system described above (light source, image sensor with IR and color pixels, processor generating IR and color pictures, skin-color identification, and IR picture filtering using skin-color data) also compares the shape of the object in the IR picture to a pre-defined shape. This shape comparison can be used for preliminary object classification or to focus the skin-color identification.
5. The detection system of claim 4 , wherein the processor is configured to compare grey scale intensity of pixels in the IR picture with a threshold to determine the object shape.
In the detection system described above (light source, image sensor with IR and color pixels, processor generating IR and color pictures, skin-color identification, IR picture filtering using skin-color data, and object shape comparison), the system determines the object's shape by analyzing the grayscale intensity of pixels in the IR picture and comparing these intensities to a set threshold.
6. The detection system of claim 1 , wherein the image sensor is configured to output at least one first picture when the light source is turned on and to output at least one second picture when the light source is turned off; and the processor is further configured to compare the at least one first picture with the at least one second picture to generate a subtraction picture to be served as the picture for generating the IR picture and the color picture.
The detection system described above (light source, image sensor with IR and color pixels, processor generating IR and color pictures, skin-color identification, and IR picture filtering using skin-color data) improves image quality by capturing pictures with the light source on and off. The processor then subtracts the picture taken with the light off from the picture taken with the light on, creating a subtraction picture used to generate the IR and color pictures, thereby reducing background noise.
7. The detection system of claim 1 , wherein the image sensor is configured to output one first picture and two second pictures, and the processor is configured to subtract an average of the two second pictures from the one first picture to generate a subtraction picture to be served as the picture for generating the IR picture and the color picture.
The detection system described above (light source, image sensor with IR and color pixels, processor generating IR and color pictures, skin-color identification, and IR picture filtering using skin-color data) captures one picture with the light source on and two pictures with the light source off. The processor averages the two "light off" pictures and subtracts that average from the "light on" picture, generating a subtraction picture used for creating the IR and color pictures.
8. The detection system of claim 1 , wherein the image sensor is configured to output two first pictures and one second picture, wherein the processor is configured to subtract the one second picture from an average of the two first pictures to generate a subtraction picture to be served as the picture for generating the IR picture and the color picture.
The detection system described above (light source, image sensor with IR and color pixels, processor generating IR and color pictures, skin-color identification, and IR picture filtering using skin-color data) captures two pictures with the light source on and one picture with the light source off. The processor averages the two "light on" pictures, then subtracts the "light off" picture from the average, generating a subtraction picture used to create the IR and color pictures.
9. A gesture recognition method adapted to a detection system, the detection system comprising an image sensor, which comprises a sensor array having both infrared pixels and color pixels, and a light source configured to illuminate an object, the gesture recognition method comprising: outputting a picture with the image sensor; generating an IR picture and a color picture according to the picture, wherein the IR picture contains an IR object image and the color picture contains a color object image which corresponds to the IR object image in the IR picture; identifying an object shape of the IR object image in the IR picture; identifying a skin-color object from the color object image in the color picture when the object shape is not a predetermined shape; filtering the IR picture with the identified skin-color object by comparing the identified skin-color object and the IR object image in the IR picture to remove an interference image, which does not correspond to the skin-color object of the color object image, from the IR object image and generate a filtered IR picture; and performing gesture recognition according to the filtered IR picture.
A gesture recognition method uses a detection system with an image sensor (IR and color pixels) and a light source. The method outputs a picture, generates IR and color pictures, identifies the object's shape in the IR picture. If the object shape is not a pre-determined shape, it identifies skin-colored areas in the color picture and filters the IR picture using this skin-color data to remove interference. Finally, it performs gesture recognition based on the filtered IR picture.
10. The gesture recognition method of claim 9 , further comprising: performing gesture recognition according to the IR picture when the object shape is the predetermined shape.
The gesture recognition method described above (image sensor with IR and color pixels, light source, generating IR and color pictures, shape identification, skin-color filtering when shape is not predetermined, gesture recognition) also performs gesture recognition directly on the unfiltered IR picture if the identified object shape matches a pre-determined shape.
11. The gesture recognition method of claim 9 , further comprising: comparing grey scale intensity of pixels in the IR picture with a threshold to determine the object shape.
In the gesture recognition method described above (image sensor with IR and color pixels, light source, generating IR and color pictures, shape identification, skin-color filtering when shape is not predetermined, gesture recognition), the object shape is determined by comparing the grayscale intensity of pixels in the IR picture against a specified threshold.
12. The gesture recognition method of claim 9 , wherein the skin-color object is identified using a skin-color algorithm based on RGB color mode, HSV color mode, YUV color mode, YIQ color mode or YCbCr color mode.
In the gesture recognition method described above (image sensor with IR and color pixels, light source, generating IR and color pictures, shape identification, skin-color filtering when shape is not predetermined, gesture recognition), the skin-color object is identified using a skin-color detection algorithm based on color models such as RGB, HSV, YUV, YIQ, or YCbCr.
13. The gesture recognition method of claim 9 , wherein the filtering is to perform an AND operation between the IR picture and the color picture as comparing the identified skin-color object and the IR object image.
In the gesture recognition method described above (image sensor with IR and color pixels, light source, generating IR and color pictures, shape identification, skin-color filtering when shape is not predetermined, gesture recognition), the filtering of the IR picture is performed by using the identified skin-color areas in the color picture as a mask for the IR picture through an AND operation.
14. The gesture recognition method of claim 9 , wherein the image sensor is configured to output at least one first picture when the light source is turned on and to output at least one second picture when the light source is turned off, and the gesture recognition method further comprises: comparing the at least one first picture with the at least one second picture to generate a subtraction picture to be served as the picture for generating the IR picture and the color picture.
The gesture recognition method described above (image sensor with IR and color pixels, light source, generating IR and color pictures, shape identification, skin-color filtering when shape is not predetermined, gesture recognition) also includes capturing pictures with the light source on and off. The system then subtracts the picture taken with the light off from the picture taken with the light on to create a subtraction picture used for generating IR and color pictures.
15. The gesture recognition method of claim 9 , wherein the image sensor is configured to output one first picture and two second pictures, and the gesture recognition method further comprises: subtracting an average of the two second pictures from the one first picture to generate a subtraction picture to be served as the picture for generating the IR picture and the color picture.
The gesture recognition method described above (image sensor with IR and color pixels, light source, generating IR and color pictures, shape identification, skin-color filtering when shape is not predetermined, gesture recognition) captures one picture with the light source on and two pictures with the light source off. The method averages the two "light off" pictures and subtracts that average from the "light on" picture, generating a subtraction picture to be used for creating the IR and color pictures.
16. The gesture recognition method of claim 9 , wherein the image sensor is configured to output two first pictures and one second picture, and the gesture recognition method further comprises: subtracting the one second picture from an average of the two first pictures to generate a subtraction picture to be served as the picture for generating the IR picture and the color picture.
The gesture recognition method described above (image sensor with IR and color pixels, light source, generating IR and color pictures, shape identification, skin-color filtering when shape is not predetermined, gesture recognition) captures two pictures with the light source on and one picture with the light source off. The method averages the two "light on" pictures, then subtracts the "light off" picture from the average, generating a subtraction picture for creating the IR and color pictures.
17. A detection system configured to detect an object, the detection system comprising: an image sensor comprising both of a plurality of infrared pixels and a plurality of color pixels, and configured to capture picture information of the object; and a processor configured to separate the picture information from the plurality of infrared pixels and the plurality of color pixels, wherein the picture information from the infrared pixels contains an JR object image and the picture information from the color pixels contains a color object image which corresponds to the IR object image in the picture information from the infrared pixels, indicate skin-color pixels among the color object image, determine an object image from the IR object image in the picture information from the infrared pixels by comparing the indicated skin-color pixels and the JR object image to remove an interference image, which does not correspond to the skin-color pixels of the color object image, from the IR object image, and identify an object motion according to the determined object image in the picture information from the plurality of infrared pixels.
A detection system identifies object motion using infrared (IR) and color data. It includes an image sensor with both IR and color pixels to capture picture information and a processor. The processor separates the IR and color data. It identifies skin-colored pixels in the color data and uses this skin-color information to filter the IR data, removing irrelevant background and interference that doesn't correspond to skin. The system then identifies object motion based on the filtered IR data.
18. The detection system of claim 17 , further comprising a light source configured to illuminate the object.
The detection system described above (image sensor with IR and color pixels, processor separating IR/Color data, skin-color identification, and IR data filtering) also includes a light source to illuminate the object being detected.
19. The detection system of claim 18 , wherein the processor is further configured to perform subtraction between the picture information from the plurality of infrared pixels obtained when the light source is turned on and turned off respectively.
In the detection system described above (image sensor with IR and color pixels, processor separating IR/Color data, skin-color identification, IR data filtering, and light source), the processor also performs a subtraction between IR data captured with the light source turned on and IR data captured with the light source turned off. This subtraction helps to reduce background noise and improve object detection.
Unknown
December 26, 2017
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.